Are Remittances a Stabilizing Factor

8/12/2019 Are Remittances a Stabilizing Factor

1/17

Are remittances a stabilizing factor in the Mexican economy?

J ORGE E DUARDO M ENDOZA C OTA 1

n Abstract: This paper evaluates whether remittances can function as a countercycli-cal mechanism for a recessive phase of the business cycle in Mexico. Remittanceshave resulted from the intense migration of Mexican workers to the USA during thenineties. The ow of remittances in Mexico had an explosive growth until the year2006; since 2007, a drastic drop in this growth is observed, and in 2008 it becamenegative. A panel data model was established using information of the GDP andremittances at the state level for the period 2005-2008. This methodology considersthe regional effects of remittances on the economic growth of the Mexican economy.The econometric results of the analysis indicated that remittances have a positivecoefcient with respect to the position of the business cycle in Mexico. Additionally,the ndings weaken the positions of those that consider the strategy of promotingremittances for use as a tool for macroeconomic stabilization of the business cycle.

n Resumen: El trabajo evala el papel de las remesas como un mecanismo contracclico

para la fase recesiva del ciclo econmico en Mxico. Los ujos de remesas derivande la intensa migracin de trabajadores mexicanos hacia los EUA durante la dcadade los noventa. El ujo de remesas creci explosivamente hasta el 2006, cayendoen el 2007 y convirtindose en negativo en 2008. En el artculo se establece unmodelo economtrico de panel utilizando el PIB y a las remesas al nivel estatalpara el periodo 2005-2008. Los resultados economtricos muestran un coecientepositivo para las remesas con relacin al comportamiento del componente cclicodel PIB en Mxico. Dichos resultados debilitan loa planteamientos respecto a laconsideracin de los ingresos por remesas como un instrumento de estabilizacinmacroeconmico en la fase recesiva del ciclo econmico.

n Key words : remittances, Mexican economy, GDP, economic cycles, economicintegration.

n JEL Clasicacin : F24, F15, F44, F36.

n Recepcin: 09/05/2011 Aceptacin: 27/09/2011

1 Researcher in the Economics Department of El Colegio de la Frontera Norte (College of the Nor-thern Border), National Researcher in Mexico, Level III. This article was funded by the NationalCouncil of Technology and Science, project 98763.


2/17

84 n EconoQuantum Vol. 9. Nm. 1

n Introduction

The Mexican economy has shown an increasing vulnerability to the recessionary stagesof the US business cycle, most recently during the economic recessions of 2001 and2008. Traditionally, external trade and foreign direct investment are the mechanisms bywhich an economic downturn in the US affects the economic activity in Mexico. Addi-tionally, as a result of the integration of the US and Mexican economies, and the intensemigration of Mexican workers to the USA, during the decade of the nineties, the remit-tances sent by Mexican workers from the US have increased rapidly, especially duringthe nineties. Therefore, when the US economy entered a recessionary phase in 2008,those channels of economic integration became a negative influence on the macroeco-nomic performance of the Mexican economy.

It is important to note that remittances have become an important source of for-eign exchange in Mexico, and have played the role of a financing mechanism thathas supplemented private foreign financing, with the advantage that these resourcesdo not create debt. Concerning this point, several papers have argued that remittancesmay function as a countercyclical stabilizer in receiving countries (Frankel, 2009). Ac-cording to this perspective, remittances can be used for smoothing the economicfluctuations of the business cycle by diversifying the financial structure and by gen-erating high return investment opportunities. Also, it has been pointed out that sinceremittances are determined by decentralized private decisions, as compared to public

spending, there are better conditions for determining the best use of remittances. There-fore, remittances could work as a countercyclical tool, which would increase when thereceiving country enters a recessionary phase and would decrease when the country isgrowing above its potential level of income. This aspect is particularly important forimplementing public policies looking to organize and make better use of remittances indeveloping countries.

From this perspective, it has been considered that remittances positively affect eco-nomic growth because they can encourage the expansion of the rate of investment andconsumption. Within a context of financial constraints, remittances could turn into amechanism aimed at funding the lack of domestic credit and therefore could furtherincrease consumption and investment.

However, in developing economies with a higher degree of economic integrationwith respect to developed economies, the synchronization of the two economies limitsthe possibility that remittances could become a mechanism for smoothing the businesscycle fluctuations in the short run. In fact, when the countries sending and receivingremittances experience economic integration, such as the case of the US and Mexico,remittances would not necessarily work as a countercyclical mechanism. The high levelof synchronization between two countries reduces the flow of remittances when thoseeconomies are facing the recessive phase of the business cycle. Because of this, the flow

of remittances become procyclical, following the same pattern as exports and FDI.Considering this fact, the objective of this paper is to evaluate the impact that mac-


3/17

Are remittances a stabilizing factor... n 85

roeconomic fluctuations in Mexico have on the flow of remittances, in order to cor-roborate the capacity of those financial resources to work as a countercyclical mecha-nism during a recessive phase of the business cycle in Mexico. The article is structured

as follows: the following section is a discussion about the effect of remittances oneconomic growth; the third section is an analysis of the macroeconomic variables, re-mittances and Mexican migration flows; the fourth section includes the theoretical ap-proach and the methodology of analysis; the fifth section includes the results of theanalysis and the last section presents the conclusions of the paper.

n Mexican migration and macroeconomic behavior

From a macroeconomic perspective, the important issues regarding remittances are theamount of remittances sent to the recipients in the home country, the income earned andbrought back to the sending country by return migrants, the mechanisms for distribut-ing the financial resources between consumption (Rempell y Lodbell, 1978) and invest-ment (Stark, 1978 y 1991), and the short and long term effects on economic growth andincome distribution.

With respect to the impact of remittances in the short run, a central aspect of theanalysis has to do with the study of remittances as a source of foreign exchange and,therefore, its effect on macroeconomic equilibrium and GDP growth. This approach isbased on the Keynesian model in order to capture the effect of remittances on the aggre-gate demand, by estimating the income multiplier. From this perspective, several stud-

ies regarding the effect of remittances on consumption, investment and imports, haveconcluded that remittances have positively affected GDP and the marginal propensityto import (Glystos, 1999, and El-Sakka and Mcnabb, 1999).

On the other hand, remittances could negatively influence economic growth,through the effect of the Dutch disease, which could result from the appreciationof the real exchange rate. With respect to this point, Acosta, Lartey and Mandelman(2007), using a VAR model, found evidence of the Dutch disease impact. Addition-ally, there is a possibility of moral hazard due to asymmetric information generated bythe distance between the sending and receiving countries. This situation determines thatindividuals receiving remittances could have fewer incentives to work or to make highrisk investments (Chami, Fullenkamp and Jahjah, 2003), thus affecting the economicgrowth of the receiving country.

A relevant aspect of the link between remittances and economic growth is relatedto the business cycle. The World Bank (2006) has found evidence that remittances arepositively related to the GDP per capita of the country sending migrants, and thus receiv-ing remittances, and a weak correlation with the GDP per capita of the country receivingmigration flows and sending remittances. In addition, Sayan (2006) found that in twelvedeveloping economies there was no evidence of the countercyclical effect of remittances.Furthermore, Lueth y Ruiz-Arranz (2007) estimated a vector correction model in order to

determine the macroeconomic impacts of remittances and found that remittances are pro-cyclical, in that they decline when the local currency weakens and increase with oil price


4/17


shocks that affect migration flows from the oil producing countries.The long run effects of remittances are related to their impact on productivity, in-

equality (Chami, Fullenkamp y Jahjah, 2003), human capital formation (Hanson and

Woodruff, 2002), and entrepreneurial activity (Woodruff and Zenteno, 2001).It is worth mentioning that the possibility that remittances could have a positiveimpact depends on circumstances such as the mobility of capital, and the perceptionof a permanent remittances flow. Both factors could reduce the effect of remittanceson capital formation in favor of consumption. Another potential impact is related toits role in the development of human capital in the countries receiving remittances(Chami, Barajas, Cosimano, Fullenkamp, Gapen and Montiel, 2008). In this case, theentrepreneurial and labor skills and the level of income of the household members thatreceive remittances could limit the impact of remittances on human capital formationand consumption.

n Stylized facts of the business cycle and remittances in Mexico

The average annual growth of remittances between 1995 and 2009 was 11.7%. In par-ticular, the annual average growth rate during the period from 1995-2003 was explo-sive at 15.7% (Table 1). However, after 2007, remittances declined sharply, reaching anegative growth rate in 2008 and 2009 of -3.6% and -15.7%, respectively. Even so, thetotal amount of remittances increased from $9,814.45 million dollars in December of2002 to $21,181.1 million dollars in 2008 (Table 2). It is worth mentioning that in the

period analyzed, the share of remittances as a percentage of GDP increased from 2.3%in 2003 to 2.8% in 2009.It is worth mentioning the rising trend of remittances was reversed by the interna-

tional economic recession of 2008, which reduced the migration of Mexican workers tothe USA and increased return migration to Mexico. The growing remittance flows ex-perimented during the decade of the nineties and, particularly between 2000 and 2005,

Table 1Rate of Growth of remittances 1995-2009

Years Rate of growth Years Rate of growth1995 2004 21.88%1996 15.00% 2005 18.32%1997 15.18% 2006 17.88%1998 15.66% 2007 1.96%1999 5.02% 2008 -3.57%2000 11.22% 2009 -15.74%2001 35.34% AAGR 2009-1995 11.68%2002 10.33% AAGR1995-2003 15.66%

2003 53.25% |AAGR 2009-2003 4.89%Source: Own elaboration with data from El Banco de Mxico (Banxico). TCPA: Annual average growth rate.


5/17


are directly related to the intense migratory activity for the same period. According tothis fact, Figure 1 shows a positive trend between the migratory flows of the Mexicanworkers and growth of remittances at the state level. However, the economic recessionof 2008 marked a change in the trend of both the migratory flows and remittances, withboth trends becoming negative.

The growth of remittances brought about an expansion of foreign exchange in theMexican economy, turning remittances into an important source of financial resources,comparable with other sources such as direct foreign investment (FDI) and exports.

In fact, the share of remittances compared to FDI represented 86.5% in 2003 and in-creased to 100.7% in 2006. From peak of that year, the participation of remittances

Table 2Evolution of the flow of remittances and FDI in Mexico (current dollars)

Remittances (A) GDP (B) A/B Foreign Direct Investment (C) A/C

2003 15,040,730,000 663,751,644,242 2.27% 17,387,878,400 86.50%

2004 18,331,310,000 723,934,708,060 2.53% 26,828,800,400 68.33%

2005 21,688,700,000 810,421,268,067 2.68% 28,461,645,000 76.20%

2006 25,566,830,000 911,928,299,700 2.80% 25,378,658,500 100.74%

2007 26,068,680,000 993,291,243,154 2.62% 36,034,696,100 72.34%

2008 25,137,370,000 1,062,681,813,970 2.37% 26,747,246,800 93.98%

2009 21,181,140,000 ND ND 21,663,000,000 97.78%

2010- 10,035,000,000 ND ND 20,400,171,600 49.19%

Source: Own elaboration with data from the Balanced of Payments Statistics Banxico.

Figure 1Relation between migration flows and

remittances at the state level (2003) (%)

14.00

12.00

10.008.00

6.00

4.00

2.00

0.000.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00

R e m i t t a n c e s

Migration flows

Source: Balance of Payments Statistics. Bank of Mexico (BANXICO) and populations projections of the NationalCouncil of Population (CONAPO).


6/17


declined to 72.2% in 2007. In the first six months of 2010 remittances only represented49.2% of FDI (Table 2). This recent trend demonstrates that, in the recessive phase ofthe business cycle, remittances follow the same negative trend as other macroeconomic

variables, thus decreasing in importance as an external source of financing.In general, one of the main sources of foreign exchange for the Mexican economysince the liberalization strategy started has been the export sector, particularly manu-facturing and oil exports. During the period of 2003-2008, the share of remittances ascompared with exports corresponded to 9.3% of total exports, 63.4% of the oil exportsand 11.4% of the manufacturing exports (Table 3). Therefore, it is evident that remit-tances have become an important source of foreign income even when compared with

Table 3Remittances and exports in Mexico

(Thousands of dollars)

Remittances(A)

Total exports(B)

Oil exports(C )

Manufacturingexports (D)

A/B A/C A/D

2003 15,040,730 164,766,436 18,597,225 140,650,306 9.13% 80.88% 10.69%2004 18,331,310 187,998,555 23,663,079 157,768,214 8.00% 63.56% 9.53%2005 21,688,700 214,232,956 31,888,572 175,195,588 7.02% 47.17% 8.59%2006 25,566,830 249,925,144 39,016,849 202,751,837 6.02% 38.55% 7.42%2007 26,068,680 271,875,312 43,013,838 219,709,422 5.53% 34.97% 6.85%

2008 25,137,370 291,342,595 50,635,372 230,881,575 5.16% 29.70% 6.51%2009 21,181,140 229,783,026 30,910,823 189,698,395 6.55% 48.66% 7.93%

2010-06 10,035,000 141,262,022 19,381,992 115,733,427 10.65% 77.60% 13.00%Average

2003-2010163,049,760 1,751,186,046 257,107,750 1,432,388,764 9.31% 63.42% 11.38%

Source: Own elaboration with data from the Balance of Payments Statistics from Banxico and the EconomicInformation Bank from INEGI.

the traditional sources of foreign exchange for the Mexican economy.The evolution of remittances at the state level illustrate that states with high migra-

tory intensity also have a high participation of remittances in the state GDP. Such is thecase of the states of Michoacan and Zacatecas, where remittances represented 10.7%and 8.1% of each states respective GDP in 2003. In addition, there are states that re-ceived a high level of remittances, even though their share of remittances to GDP islower due to a high level of economic activity. The states of Veracruz, Puebla, State ofMexico and the Federal District fall into this category and, in 2008, remittance flowsare of $1,620, $1,568, $1,942 and $1,105 million dollars received, respectively. How-ever, in those states, remittances only represented 3.3%, 4.4%, 2.1% and 0.6% of GDP(Table 4). Finally, it is worth mentioning the accelerated rate of growth of remittances

in the states with extensive migration of Mexican workers to the USA, such as Chiapas,Oaxaca, Veracruz and Hidalgo. Therefore, the empirical analysis of the stylized facts


7/17


Table 4Remittances and GDP In Mexico, 2006-2008

Remittances (Millions of dollars) GDP (%) 2003 2008 2003 2008Total 15,040.7 25,137 2.27 2.37Aguascalientes 260.9 $332 3.66 3.04Baja California 144.4 $342 0.72 1.15Baja California Sur 19.4 $36 0.55 0.59Campeche 52.5 $74 0.16 0.10Coahuila de Zaragoza 142.2 300 0.65 0.90Colima 105.2 198 2.82 3.62Chiapas 439.3 800 3.34 4.16

Chihuahua 240.5 475 1.09 1.43Distrito Federal 826.8 1,105 0.67 0.61Durango 265.3 450 3.03 3.55Guanajuato 1,112.1 2,096 4.14 5.33Guerrero 1,403.2 2,325 13.01 15.42Hidalgo 845.5 1,402 8.96 8.12Jalisco 589.1 $940 1.32 1.44Mxico 1,345.4 1,942 2.25 2.08Michoacn de Ocampo 1,778.9 2,457 10.73 9.52

Morelos 368.5 621 4.40 5.74Nayarit 229.6 384 5.95 6.13Nuevo Len 193.3 331 0.40 0.41Oaxaca 770.8 1,457 7.31 8.73Puebla 804.9 1,568 3.55 4.43Quertaro 283.2 442 2.59 2.30Quintana Roo 53.7 100 0.59 0.66San Luis Potos 397.7 758 3.34 3.86Sinaloa 319.4 $489 2.36 2.27Sonora 130.5 $318 0.83 1.23Tabasco 87.3 $159 0.55 0.41Tamaulipas 238.1 $511 1.05 1.40Tlaxcala 143.1 $299 3.68 5.43Veracruz de Ignacio de la Llave 989.6 1,620 3.41 3.32Yucatn 59.5 $129 0.67 0.90Zacatecas 400.5 $678 8.09 8.28Standar deviation 453 700 3.22 3.44Average 470 786 3.31 3.64

Source: Own elaboration based on the statistics of the National Accounting System from INEGI and Banxico.


8/17


of remittances behavior with respect to the macroeconomic variables in the Mexicaneconomy reveals that, for the period 2003-2008, remittances have followed the trend ofthe expansive and recessive phases of the major macroeconomic variables.

With the objective of understanding the role of remittances as a source of financialfunds that affect the behavior of the business cycle, this paper estimates the correlationof the cyclical components of the GDP and remittances for the period from January1995 to February 2010. For that purpose, the Hodick-Prescott filter (HP) was used toseparate the cyclical component and the trend of economic time series. The Hodrick-Prescott method is a straightforward technique where the outcome of the estimationdoes not require fine tuning.

However, this method has been criticized because of several underlying limitationssuch as the discretional choice of the smoothing parameter . In addition, any trend esti-mation procedure is confronted with the problem that the trend of a macroeconomic timeseries is not observable. This makes it difficult to directly assess the quality of a specificestimate of the trend component. In order to offset that uncertainty, the decomposition ofa time series into a trend and a cyclical component requires the identification of assump-tions regarding the functional form and stochastic properties of the trend component.

There are other procedures such as the band pass filter, developed by Baxter andKing (1995). However, it has similar properties to the HP filter, such that this filter isalso symmetric, the filter weights are fixed and the resulting moving average filter is ofinfinite order. Therefore, so far, there is not an ideal filter to calculate business cycles.

The Hodrick-Prescott method considers that a series Y i consists of a cyclical compo-

nent ( C ) and a trend ( Yi = Ti + Ci ). By applying the HP filter it is possible to determinethe conditions of the times series; the first term follows a smooth trend and the secondone follows the short term movements of the original series. The trend is estimatedwith a process of optimization. 2 The way to control between the sensitivity and thesmoothness is by using a parameter > , which defines one of the two conditions. Forhigh values, the smoothness condition predominates and for low values, the adjustmentcondition takes over.

Figure 2 shows the cyclical component and the trend for both the remittances andthe GDP series. It stands out that the peaks of expansion and the downturns of bothseries exhibit similar movements, visually supporting the assumption of a positive cor-relation. Moreover, it can also be deduced from the graph that there is a higher volatilityin the cyclical component of remittances as compared to that of the GDP.

The analysis of the times series with the trend shows that there is a positive correla-tion between the two series and, therefore, the flow of remittances is procyclical withrespect to the movement of the GDP with three quarterly lags. This correlation givesevidence that, in a context of economic integration between the USA and Mexico,Mexican remittances are characterized by a positive movement with respect to the be-

2 min T T T T T T Y

t t t t t t t

t

T

t

T 2

1 1 22

313

m- - - -- - -= ^ ^ ^h h h6; @ E//

subject to Y t = T t + C t , where the rst term related the

accuracy of the trend with the original series and the second one the degree of smoothness of the trend.


9/17


Figure 2Mexico: index of growth of remittances and GDP (Dollars)

Figure 3Cyclical component of the GDP and remittances series: 2005-2010

Table 5Estimation of correlation and volatility of quarterly series of GDP and Remittances,

1993-01-2010-12

Correlation (1) 0.169



Autocorrelation (1) 0.765



SD PIB 0.023

SD REM 0.078

Volatility 3.36Source: Own estimations with data from Banxico and INEGI. SD= Standard deviation

Source: Own elaboration based on the statistics of the National Accounting System from INEGI and BANXICO.

140

120

2003 2004 2005 2006 2007 2008

100

80

60

40

20

0

GDP

Remittances

Source: Balance of Payments Statistics. Banco de Mxico.

1 9 9 5 Q 1

1 9 9 6 Q 1

1 9 9 7 Q 1

1 9 9 8 Q 1

1 9 9 9 Q 1

2 0 0 0 Q 1

2 0 0 1 Q 1

2 0 0 2 Q 1

2 0 0 3 Q 1

2 0 0 4 Q 1

2 0 0 5 Q 1

2 0 0 6 Q 1

2 0 0 7 Q 1

2 0 0 8 Q 1

2 0 0 9 Q 1

2 0 1 0 Q 1

CCPIB

CCREM

0.15

0

0.1

0.05

-0.05

-0.1

-0.15

-0.2


10/17


havior of the Mexican GDP (Table 5). Another aspect that stands out regarding thesetwo variables has to do with the high level of volatility of the cyclical component ofremittances when compared to that of the GDP, as indicated by a volatility index

of 3.36, far above the unity.

n Theoretical and methodological

Theoretical approachThe approach to understanding the relationship between remittances and the GDP isderived from a production function:

, ,Y AK L 0 1 1t t t 1 1a b a= = -a b

Where A is the technology, K and L are capital and labor, is the income share ofcapital and is the share of labor in the period t .

The economy functions with income generated in the domestic economy and withthe income from remittances, according to the following identity:

(1) C I G Y R+ + = +

Where C is the value of aggregate consumed goods, I is the level of aggregate in-vestment, G is government spending, Y is the national income and R are remittances.

In other words, income and remittances could be used for consumption and investmentin the economy. This model assumes that there are linear functions for consumptionand investment, the government spending ( G) is an exogenous variable, imports ( M )depend on disposable income and exports ( X ) depend on the disposable income in theUS. The functions that show the behavior of these variables are:

(2) C Y Rdmxa b d= + +^ ^h h

(3) , I I r h= ^ h

(4) M M R0 c= + ^ h

(5) X X Y deua0 c= + ^ h

Where a , M o y X o are autonomous parameters, and d and c are parameters show-ing the effect of remittances on consumption and imports, all of them are positive andbetween 0 and 1; r and h are the interest and the investor confidence parameters.Therefore, remittances are incorporated in the theoretical analysis of the GDP determi-nation as a variable that impacts the patterns of consumption and imports.

Additionally, it is assumed that remittances are procyclical in regards to the GDP:

(6) RY

Y t

t z=m

c m


11/17


Where Y t is the actual GDP growth, Y is the potential growth of GDP, and z y

are positive and constant parameters.

Taking the logs of equation (6) we obtain:

(7) ln ln RY

Y t

t z m= -+ c mEquation 7 is the basis of the econometric estimation of the effect of the movements

of GDP on the remittance flows to Mexico. Hence, changes in the proportion betweenactual and potential GDP could show a positive or negative correlation with respect tothe movements of remittances. The flow of remittances adds a new macroeconomicmechanism, which may have the potential to work as an automatic stabilizer. Someauthors have founded that remittances are relatively stable when compared with fi-nancial flows (Chami, Hakura and Montiel, 2009), and therefore providing a basis forassuming that remittances could work as a countercyclical mechanism. However, thisassumption depends on the effect of the reduction of the GDP and income on the labormarkets. Hence, the final effect of remittances on the macroeconomic stabilization isambiguous, depending on the economic conditions of the receiving country, and mak-ing it relevant to estimate whether or not remittances are procyclical or countercyclicalfor the case of the Mexican economy in the context of its economic integration withthe U.S. economy.

n Methodological aspects

In order to analyze the relationship between remittances and the GDP at the state level,a panel data model was established using state GDP and remittances data for the period2005-2008. This methodology is utilized because its approach allows taking into consider-ation the regional effects of remittances on the economic growth of the Mexican economy.

The objective of the paper is to corroborate the existence of the correlation of thecyclical component of the GDP and the flows of remittances at the Mexican state level.The methodology consists of a regression model that includes as a potential shockvariable, the cyclical component of the Mexican GDP (macroeconomic determinant)and the cyclical component of the U.S. GDP (GDP of the host country). The stock ofmigrants by state is included as a control variable, since the cyclical behavior of re-mittances is probably related to the stock of migrants (Frankel, 2009). Additionally,income per capita by states in Mexico is incorporated as a variable representing theeffect of regional purchasing power and consumption disparities on the decisions thatdetermine the regional flows of remittances. According to Ramrez and Sharma (2008),the economic impact of remittances tends to be higher in the regions and groups withhigher income. Finally, remittances behave differently than private capital flows and

apparently are less volatile than FDI (Fullenkamp, Gapen and Montiel, 2008). There-fore, FDI is included in the model to corroborate whether or not there is a correlation


12/17


between remittances and foreign exchange flows. Thus, based on equation (7), the econometric equation of the empirical research is

constructed as follows:

ln R B M GDPpc CC GDP FDI it i it it it

it it it 0 1 1 2 2 3 4 4 b m b b b y= + + + + + +^ ch m

i = 1....32, t = 4

Where:im = individual effect of the error term ( it = i+ it)

GDP it /GDPtrend it = GDP of state i divided by trend component of the GDP at the statelevel (cyclical position o the GDP of state i at time t ),M = stock of Mexican migrants in the USA,GDPpc = GDP per capita,R = Remittances at the state level,FDI = Foreign direct investment divided by GDP at the state level

This type of methodology allows the estimation of either a model of fixed effects ora model of random effects, according to the results of the Hausman and Breuch-Pagantests. The data base used in the econometric estimation consists of an annual panel dataset for 31 states and the Federal District of Mexico for the period 2005-2008. For theremittances variable the information was obtained from the Balance of Payments Sta-

tistics published by the Bank of Mexico, to calculate the proxy of the stock of Mexicanmigrants by state, information from the Survey of Migration of the Northern Border(EMIF) was used, for the GDP at the state level the National Accounts Statistics pub-lished by the National Institute of Statistics, Geography and Informatics (INEGI) wasconsulted, the GDP per capita at the state level was constructed using the populationprojections of the National Council of Population (CONAPO) and the national accountdata from INEGI. Finally, the information on foreign direct investment was obtainedfrom INEGI.

n Analysis of results

The purpose of the panel data model was to determine whether or not remittancesfrom Mexican migrants are procyclical. That is, whether remittance flows increase ordecrease in the same direction as the GDP or if they exhibit a movement in the oppositedirection with respect to the business cycle.

The panel data model was analyzed with both fixed and random effects. Both thelevel and the cyclical component of remittances were used as dependent variables withthe aim of capturing a better estimate of the synchronization of that variable to theMexican GDP cyclical component. A set of estimations was conducted by combining

the cyclical component of the Mexican GDP variable with a group of control variables.In order to determine a more suitable econometric estimation regarding the possibil-


13/17


ity of non-observable heterogeneity of the regional sample that could affect the econo-metric results, a Hausman test was performed to corroborate the possible inconsistencyof the results of the random effects model. Finally, the results of the Breusch-Pagan test

rejected the null hypothesis for the migration stock variable in the model that includesremittances in levels as the dependent variable. In the case of the model that includesthe cyclical component of remittances as the dependent variable, three estimations re-

jected the null hypothesis. From the results of the estimation of the fixed effects model, the following aspects

stand out: the coefficient of the cyclical component of the GDP showed a positive correla-tion with the coefficient of remittances and was statistically significant when it was esti-mated with migration and the Mexican per capita GDP (Table 6). This panel data by statesestimation corroborates the results of the analysis of the cyclical component estimated byapplying the Hodrick-Prescott filter to remittances and GDP at the national level.

When substituting the cyclical component of remittances as a dependent variableand including the stocks of Mexican migrants in the USA, the cyclical component ofthe Mexican GDP showed a positive and statistically significant coefficient. It is worthmentioning that neither the cyclical component of the U.S. GDP and the FDI exhibiteda statistically significant coefficient although they were both positive.

Although the estimation with remittances in levels as the dependent variable did notreject the Breach Pagan test, a random effect model was estimated for all the modelsincluding control variables (Table 7). The estimations showed positive and statisticallysignificant coefficients for the cyclical component of the Mexican GDP at the 5% and

10% confidence level. However, for the case of the cyclical component of remittancesas the dependent variable, the coefficients were statistically insignificant. However, thecontrol variable of the U.S. component cycle presented a positive and statistically sig-nificant coefficient. Since there is evidence of the synchronization of the GDP cyclicalcomponents between the U.S. and Mexico; this result further supports the evidence ofa pro-cyclical relationship between remittances and the economic activity in Mexico.

With regard to the other control variables of the model, the estimation displayed thecoefficient of the GDP per capita with a positive sign and it was statistically significantfor the case of remittances as the independent variable in fixed and random effectsmodels, but when using the cyclical component of remittances, the results were notstatistically significant. Therefore, the results are not conclusive.

Finally, the coefficient of the IED was positive, although its magnitude was rathersmall and was not statically significant, suggesting the possibility that both remittancesand IED have been impacted in the same fashion by the increasing economic integra-tion between Mexico and the USA. This trend reveals a positive relationship betweenfinancial flows and the GDP movement.

Therefore, the results provide empirical evidence that does not support the resultspresented by Frenkel (2009), which sustained the hypothesis that remittances could beconsidered as an instrument for smoothing the movements of the business cycle. On

the contrary, the results suggest that there is an inverse correlation between the statewith lower per capita income and the remittances flows received in Mexico at the state


14/17


T a b

l e 6

F i x e

d e f

f e c t s p a n e l m o d e l

( 2 0 0 5 - 2 0 0 8 )

3 2 c r o s s s e c t

i o n s a n

d 4 t i m e s s e r i e s

D e p e n

d e n t v a r i a b

l e : r e m

i t t a n c e s a t

t h e s t a t e

l e v e l

D e p e n

d e n t v a r i a b

l e : c y c l

i c a l c o m p o n e n t o f r e m

i t t a n c e s a t

t h e s t a t e

l e v e

l

C o n s t a n t

6 . 1

9 6 * * *

6 . 1

9 5 * * *

- 9 . 0

4 3 * *

- 1 1 . 0 5 *

- 1 1 . 0 5 8

0 . 0 3 7 * * *

0 . 0 3 * * *

1 2 . 6

7 1 * * *

- 0 . 3

5 4

- 0 . 3

5 5

( 6 . 7

9 )

( 5 4 0

. 0 0 )

- ( 2

. 3 8 3 )

- ( 1

. 8 2 7 )

- ( 1 . 8 2 7 )

( 1 5 6

. 7 2 )

( 1 7 . 4 3 )

( 3 . 2

7 )

- ( 0

. 1 6 7 )

- ( 0

. 1 6 7 )

C C P I B M X

1 . 6 2 1 * *

1 . 5 2 3 * * *

0 . 2 8 6

0 . 0 4 7

0 . 0

3 6

0 . 5 4 1 *

0 . 1 8 4

1 . 2 1 *

- 0 . 3

4 4

- 0 . 3

4 4

( 3 . 5

1 )

( 3 . 1

6 )

( 0 . 4

1 )

( 0 . 0

5 )

( 0 . 3

2 )

( 1 . 8

9 )

( 0 . 5

0 )

( 1 . 8

3 )

- ( 0

. 8 5 )

- ( 0

. 8 5 )

M i g

0 . 0 0 1

0 . 0

0 2 * * *

0 . 0 0 2 * * *

0 . 0 0 2 * * *

0 . 0 0 4 * * *

0 . 0 0 2 * * *

0 . 0 0 2 * * *

( 1 . 5

2 )

( 3 . 0

1 )

( 3 . 1

0 )

( 3 . 0

8 )

( 3 . 0

3 )

( 2 . 8

3 )

( 2 . 8

3 )

P I B p c

1 . 3 7 0 * * *

1 . 5 5 1 * * *

1 . 5 5 1 * * *

- 1 . 1

3 7

0 . 0 3 3

3 . 3 9 8

( 4 . 0

2 )

( 2 . 8

5 )

( 2 . 8

2 )

- ( 3 . 2 6 2 )

- ( 0

. 1 7 )

( 1 1 . 3 9 )

C C P I B U S

0 . 5 2 3

0 . 5

1 9

3 . 3 9 * * *

0 . 0 0 0

( 0 . 7

2 )

( 0 . 7

2 )

( 1 1 . 3 2 )

- ( 0

. 0 6 )

F D I

0

0 . 0 3

( 1 . 2

5 )

- ( 0 . 3 6 )

R 2

0 . 9 9 2

0 . 9 8

0 . 9 8 4

0 . 9 9 4

0 . 9

9 4

0 . 9 7 3

0 . 9 7 5

0 . 9 8 5

0 . 9 9 1

0 . 9 9 1

A d j u s t e d

R 2

0 . 9 8 1

0 . 9 9

0 . 9 8 2

0 . 9 9 2

0 . 9

9 2

0 . 9 6 4

0 . 9 6 6

0 . 9 7 3

0 . 9 8

0 . 9 8 8

S c h w a r z

C r i

t e r i o n

- 9 7 . 2 8 4

- 0 . 9

3

- 1 2 6

. 2 7 3

- 1 2 3

. 0 4 7

- 1 1 8 . 2 9 7

- 1 0 7

. 8 2 8

- 1 1 2

. 9 3

- 1 3 7

. 7 3 7

- 2 2 4

. 5 9 8

- 2 3 9

. 7 4 7

A k a i k e C r t e r

i o n

- 1 9 1

. 4 0 1

- 1 9 0

. 1 0

- 2 2 6

. 0 9 4

- 2 2 5

. 7 2 2

- 2 2 3 . 8 2 3

- 2 0 1

. 9 4 5

- 2 0 9

. 9 3

- 2 3 7

. 5 5 8

- 3 4 7

. 2 7 1

- 3 4 5

. 2 7 2

H a n n a - Q u i n n

- 1 5 3

. 1 5 1

- 1 5 0

. 6 6 6

- 1 8 5

. 5 3 6

- 1 8 4

. 0 0 5

- 1 8 0 . 9 4 7

- 1 6 3

. 7 0 5

- 1 7 0

. 5 3

- 1 9 7

. 0 0 1

- 3 0 5

. 5 5 5

- 3 0 2

. 3 9 7

C C G D P = c y c l

i c a l c o m p o n e n t o f

G D P , M = m

i g r a n t s

f l o w

f r o m

U S A

, F D I = F o r e

i g n

d i r e c t

i n v e s t m e n

t . * * * s i g n i f

i c a n

t a t 1 %

, * * s i g n

i f i c a n

t a t 5 % , t r a

t i o s

i n p a -

r e n t

h e s i s

.

S o u r c e :

O w n e l a b o r a

t i o n

b a s e

d o n

t h e s t a t

i s t i c s o

f t h e

N a t

i o n a l

A c c o u n

t i n g

S y

s t e m

f r o m

I N E G I a n

d B a n x

i c o .


15/17


T a b

l e 7

R a n

d o m e f

f e c t s p a n e

l m o d e l

( 2 0 0 5 - 2 0 0 8 )

3 2 c r o s s s e c t

i o n s a n

d 4 t i m e s s e r i e s

D e p e n

d e n t v a r i a b

l e : r e m

i t t a n c e s a t

t h e s t a

t e l e v e

l

D e p e n

d e n t v a r i a b

l e : c y c l

i c a l c o m p o n e n t o f r e m

i t t a n c e s a t t

h e s t a t e

l e v e

l

C o n s t a n

t

6 . 1

9 6 * * *

6 . 1

9 3 * * *

1 . 8 4 5

2 . 8 3

2 . 6 7 7

0 . 0 3 7 * * *

0 . 0 3

3 . 8 9 6 * * *

0 . 4 4 9

0 . 4 4 7

( 3 2 . 1 8 )

( 3 4 . 4 2 )

- 0 . 8

3 6

- 1 . 1

9

( 1 . 1

1 8 )

( 0 . 8

0 )

( 0 . 3 2 )

( 2 . 6

5 7 )

( 0 . 3 6 4 )

( 0 . 3

5 7 )

C C P I B

M X

0 . 7 2 3 * *

0 . 6 5 3 * *

0 . 4 9 1 *

0 . 5 5 7 *

0 . 5 5 3 *

0 . 0 8 2

0 . 0 3 2

0 . 1 2 8

- 0 . 8

8

- 0 . 0

0 9

( 2 . 3

0 )

( 2 . 0

6 )

( 1 . 6

9 )

( 1 . 6

9 9 )

( 1 . 7

8 3 )

( 0 . 4

5 )

( 0 . 1 8 )

( 0 . 6

5 4 )

- ( 0

. 5 1 5 )

- ( 0

. 5 1 7 )

M i g

0 . 0 0 2

0 . 0 0 2 * *

0 . 0 0 3

- 0 . 0

0 3

0 . 0 0 5 * * *

0 . 0 0 5 * * *

0 . 0 0 2 *

0 . 0 0 2 *

( 1 . 1

9 )

( 1 . 2

6 )

( 1 . 4

8 8 )

- ( 1

. 4 9 2 )

( 2 . 9 8 )

( 3 . 1

0 7 )

1 . 7 7 4

( 1 . 7

6 9 )

P I B p c

0 . 3 9 1 *

0 . 3 0 3

0 . 3 1 7

- 0 . 3

4 7 * * *

( 0 . 0

3 9 )

- 0 . 0

3 9

( 1 . 9

8 )

- 1 . 4

2

( 1 . 4

7 5 )

- ( 2 . 6 4 2 )

( 0 . 3 5 2 )

- ( 0

. 3 5 1 )

C C P I B

U S

0 . 5 5 5

- 0

3 . 3 0 9 * * *

3 . 3 1 * * *

( 1 . 2

4 )

( 1 . 1

1 7 )

( 1 2 . 6 3 )

( 1 2 . 5 7 )

F D I

0

0

( 0 . 3

4 2 )

( 0 . 0

2 4 )

S c h w a r z

C r i

t e r i o n

3 9 1 . 8 3 9

3 9 4 . 4 5

- 1 2 6

. 2 7 3

4 2 8 . 2 6 3

4 3 4 . 2 6 5

2 0 6 . 9 2 7

2 1 0 . 4 3 8

2 3 2 , 2 5 8

2 1 8 . 9 1 7

2 2 3 . 7 9 6

A k a i k e

C r t e r

i o n

3 8 6 . 1 3 5

3 8 5 . 8 9 4

- 2 2 6

. 0 9 4

4 1 4 . 0 0 4

4 1 7 . 1 5 3

2 0 1 . 5 4 1

2 0 1 . 8 8 2

2 2 0 . 8 5

2 0 4 . 6 5 7

2 0 6 . 6 8 4

H a n n a - Q u i n n

3 8 8 . 4 5 2

3 8 9 . 3 7 1

- 1 8 5

. 5 3

4 1 9 . 7 9 5

4 2 4 . 1 0 6

2 0 3 . 5 4 1

2 0 5 . 3 5 9

2 2 5 . 4 8 5

2 1 0 . 4 5

2 1 3 . 6 3 7

B r e u s c h - P a g a n

t e s t :

C h i - s q u a r e

( 1 )

1 8 7 . 9 6 6

0

1 4 7 . 9 0 6

1 8 1 . 9 2 5

1 8 1 . 9 3 7

1 7 8 . 4 7 3

1 7 9 . 2 4 3

1 7 9 . 4 3 4

1 8 7 . 2 6 2

1 8 6 . 7 9 9

p - v a l u e =

P ( 3 1 , 1 5 6 )

0

6 . 0 2

0

0

0

0

0

0

0

0

H a u s m

a n t e s t : C

h i - s q u a r e (

4 )

5 . 5 7 9

0 . 0 4

1 7 . 7

2 4

4 8 . 4

1 5

4 7 . 9

6 9

1 . 1 2 7

0 . 2 3 5

1 7 . 7

2 4

1 . 0 0 7

1 . 1 2 9

p - v a l u e

0

0 . 7 4

0

0

0

0

0 . 8 8 5

0

0 . 9 0 8

0 . 9 5 1

C C G D P = c y c l

i c a l c o m p o n e n t o f

G D P , M = m

i g r a n t s

f l o w

f r o m

U S A

, F D I = F o r e i g n

d i r e c t

i n v e s t m e n

t . * * * s i g n

i f i c a n

t a t 1 %

, * * s i g n

i f i c a n

t a t 5 %

, t r a

t i o s

i n p a r e n t

h e s i s .

S o u r c e :

O w n e l a b o r a

t i o n

b a s e

d o n

t h e s t a t

i s t i c s o

f t h e

N a t

i o n a l

A c c o u n

t i n g S y s t e m

f r o m

I N E G I a n

d B a n x

i c o .


16/17


level. This finding is probably related to the positive sign of the coefficient of the stockof migrants, which is the same as the coefficient found by Frankel (2009) and Freundand Spatafora (2005).

n Concluding remarks

The remittance flows are an expression of the intense migration of Mexican workersto the USA in particular during the decade of the nineties. For that reason, during thatdecade, the flow of remittances to Mexico exhibited explosive growth until the year2006. However, since 2007, a drastic fall on its growth is observed, and in 2008, thetrend of remittances became negative. It is worth mentioning that the growth of remit-tances caused its share to GDP to reach 2.8% in 2006, although it fell to 2.3% in 2008.

The reduction of remittances and migration flows since 2007 has been related tothe US economic recession. Therefore, the behavior of remittances and migration isanother expression of the economic integration between Mexico and the USA, and bothtend to decrease with the synchronized recessive phase of the business cycles experi-enced in those countries.

In particular, the econometric analysis used for estimating the correlation betweenthe GDP and remittances at the state level and cyclical component of remittances, basedon a panel model for the period 2005-2008, suggests that remittances have a positivecoefficient with respect to the position of the business cycle in Mexico. Other interest-ing conclusions that can be derived from the econometric results are the following:

l The international economic recession, within a context of increasing economic in-tegration, has become an important limitation for the effectiveness of the flow ofremittances as a countercyclical tool.

l Increases in the migration of Mexican workers are directly and positively correlatedto the expansion of remittances.

l The policies to control migration of Mexican workers to the US have also becomea limitation for the remittances to work as a countercyclical tool, particularly in thelong run.

Therefore, the results of the econometric model estimated in this paper, based on adata set at the state level for a period that encompasses a recessive phase of the businesscycle in Mexico, does not corroborate previous papers that have estimated a counter-cyclical relationship between remittances and the cyclical component (Frankel, 2009).On the contrary, the results showed a procyclical position for the case of the Mexicaneconomy. Such findings weaken the positions that consider the strategy of promotingremittances in order to create conditions to use them as a tool for macroeconomic sta-bilization of the business cycle, particularly in developing economies that are linked todeveloped economies not only in terms of migration flows, but also in terms of macro-

economic variables such as exports, FDI and financial flows.n References


17/17


Acosta Pablo, Emmanuel K. K. Lartey, and Federico S. Mandelman (2007). Remittanc-

es and the Dutch Disease, Federal Reserve Bank of Atlanta , Working Paper 2007-8.

Chami, Ralph, Adolfo Barajas, Thomas Cosimano, Connel Fullenkamp, Michael Gapenand Peter Montiel (2008). Macroeconomic consequences of remittances, Occa-sional Paper No . 259 (International Monetary Fund).

Chami Ralph, Connel Fullenkamp and Samir Jahjah (2003). Are Immigrant remmitanceflows a source of capital for development?, IMF Working Paper, WP/03/189.

Frankel, A. (2009). Are bilateral remittances countercyciclal?, NBER, Working Paper 15419.

Freund, C. L. and N. Spatafora (2005). Remittances: Costs, Determinants, and Infor-mality, World Bank Policy Research Working Paper No. 3704, Washington, DC.

Fullenkamp C., M. Gapen and P. Montiel (2008). Macroeconomic consequences ofremittances, Ocassional Paper 259, International Monetary Fund.

Hanson, G. H. and C. Woodruff (2002). Emigration and educational attainment in Mex-ico, Mimeo., University of California at San Diego.

Hodrick R. J. and E. C. Prescott (1980). Post-war U.S. business cycles: an empiricalinvestigation, Carnegie-Mellon University Discussion Paper , No. 451.

Hodrick, R. and Prescott E. (1997). Post-war U.S. business cycles an empirical investi-gation, Journal of Money Credit and Banking , vol. 2, num. 1, pp. 1-16.

Holst, Elke and Mechthild Schrooten (2006). Migration and Money What determinesremittances? Evidence from Germany, DIW Berlin, German Institute of Economic

Research , Discussion Paper 556.King R. G. and S. T. Rebelo (1993). Low frequency filtering and real business cycles,

Journal of economic Dynamics and Control , 17, pp. 207-231Lueth, Erik and Marta Ruiz-Arranz (2007). Are workers remittances a hedge against

macroeconomic shocks? The case of Sri Lanka, IMF Working Paper , WP/07/22.Rapoport, Hillel and Frdric Docquier (2005). The economics of migrants remittanc-

es, Institute for the Study of Labor (IZA) , Discussion Paper No. 1531.Ramirez D. M. and Hari Sharma (2008). Remittances and Growth in Latin America: A

Panel Unit Root and Panel Cointegration Analysis, Department of Economics, YaleUniversity, Working Paper No. 51.

Rempel, H. and R. Lobdell (1978). The role of urban-rural remittances in rural develop-ment, Journal of Development Studies , 14: 324-41.

Stark, O. (1978). Economic-demographic interaction in agricultural development: thecase of rural-to-urban migration , Rome: UN Food and Agriculture Organization.

Sayan, Serdar (2006). Business Cycles and Workers Remittances: How Do MigrantWorkers Respond to Cyclical Movements of GDP at Home? IMF WP/06/52, February.

Woodruff, C. and R. Zenteno (2001). Remittances and micro-enterprises in Mexico,Mimeo., University of California at San Diego.

World Bank (2006). Economic Implications of Remittances and Migration, Global

Economic Prospects , World Bank, pp. 86-92.

Are Remittances a Stabilizing Factor

Documents

Transcript of Are Remittances a Stabilizing Factor